AI Optical Character Recognition
Function Description
This step performs intelligent recognition of characters such as letters and digits in images. Through image acquisition and target-region setup, combined with deep-learning training and validation, it outputs character-recognition results.
|
Using this step requires a "Lite AI" feature license. Contact Mech-Mind sales to obtain it. |
Usage Scenario
This step is suitable for character-recognition scenarios with complex backgrounds, multiple fonts, or irregular characters. It is typically preceded by 2D Smart Camera or 2D Matching. The former is suitable when target positions are relatively fixed, where images are acquired and then recognized. The latter is suitable when target positions are offset, where pose-transformation parameters are obtained through alignment steps and input to this step to synchronously adjust target-region positions for stable recognition.
Input and Output
Input
| Input Port | Data Type | Description |
|---|---|---|
Image |
Image |
Images containing characters to be recognized. |
2D Alignment Parameter Group |
Pose2D[] |
Synchronously adjusts poses of target regions based on pose transformations of target objects. |
Output
| Output Port | Data Type | Description |
|---|---|---|
Recognition Result |
String[] |
Character-recognition result. |
Recognition Judgment Result |
String[] |
Recognition judgment result: OK or NG. |
Recognition Status |
Bool[] |
Indicates whether recognition succeeds. |
Detected Image |
Image[] |
Images with detection results. |
Parameter Description
| Parameter | Description |
|---|---|
Model Name |
Description: Select a configured model from the drop-down list. Models need to be added in the editor in advance. Adjustment instruction: Click Configuration Wizard (or the Configuration Wizard button on the step block) to enter the AI Optical Character Recognition Tool and configure models. For usage, refer to AI Optical Character Recognition Tool. |
Visualize Detection Results |
Description: After selection, this step outputs images with detection results for debugging and result review. Default value: Not selected. |